INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surfactant deficiency is known to underlie the infant respiratory distress syndrome (IRDS), and complex disturbances of the alveolar surfactant system are suggested to be of major importance for the pathogenesis of the acute respiratory distress syndrome in the adult (ARDS) (1). Surfactant replacement therapy is currently considered for restoration of gas exchange in this disease (4). Tracheal instillation of liquid surfactant material represents the most common route of exogenous surfactant application presently employed, with its efficacy being established in several experimental models of acute lung injury (7).

Drawbacks to this approach include the fact that ( 1) the acute liquid filling of the tracheobronchial space, suitable for meniscus formation and homogeneous peripheral transport of the surfactant material (8) may create acute deterioration of gas exchange because of airway occlusion, and (2) large amounts of surfactant are required for significant improvement of arterial oxygenation. In experimental models of acute lung injury and under clinical conditions of ARDS, surfactant doses in the range of 50 to 500 mg phospholipids per kg body weight (bw) are employed, which is remarkably high in view of the fact that the physiologic alveolar surfactant pool is estimated to approximate 10 mg/kg bw (11).

Against this background it is of interest that Lewis and coworkers (14), in a series of studies including different models of acute lung injury, demonstrated marked improvement of arterial oxygenation upon jet nebulization of comparably low quantities of surfactant, resulting in lung deposition of 4 to 5 mg/kg bw over a 3-h aerosolization period. Occasional superiority over tracheal instillation of manifold higher surfactant doses was observed; however, dependency on the type of lung injury, e.g., homogeneous versus heterogeneous lesions, was noted at the same time. Henry and coworkers (18) and this laboratory (13) have recently shown that ultrasonic nebulization may also be employed for delivery of natural liquid surfactant, allowing more efficient transtracheal aerosol transport of this material. In a model of acute lung injury by lavage maneuvers, we now compare the impact of low quantities of ultrasonically delivered versus conventional quantities of tracheal-instilled surfactant on ventilation-perfusion matching, as assessed by the multiple inert gas elimination technique. We hypothesized that both techniques might be appropriate to reduce shunt flow, but that the tracheal filling with large quantities of liquid surfactant material might create more pronounced ventilation-perfusion (A/) mismatch because of ventilation inhomogeneities as compared with the ultrasonically delivered material.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials

Alveofact was kindly provided by Dr. Karl Thomae GmbH (Biberach an der Riss, Germany). Halothane was supplied by Hoechst AG (Frankfurt am Main, Germany). All other analytical grade biochemicals, including diethyl ether and acetone, were purchased from Merck (Darmstadt, Germany). Gas mixtures of sulfur hexafluoride (SF ₆), ethane, and cyclopropane (10, 20, and 70%) were from Messer Griesheim (Siegen, Germany). The ultrasonic nebulizer (Pulmo Sonic 5500; DeVilbiss Medizinische Produkte GmbH, Langen, Germany) was a gift from DeVilbiss.

Isolated Lung Model

The perfused lung model has been previously described (19). Briefly, rabbits of either sex weighing 2.5 to 3.1 kg were deeply anesthetized and anticoagulated with heparin (1,000 U/kg). After tracheostomy the animals were ventilated with room air, using a Harvard respirator (cat/rabbit ventilator; Hugo Sachs Elektronik, March Hugstetten, Germany). After midsternal thoracotomy, catheters were placed into the pulmonary artery and the left atrium, and perfusion with Krebs-Henseleit buffer, containing 4% wt/vol hydroxyethylamylopectine (KHHB), was started. In a recirculating system the flow was slowly increased to 120 ml/min (total volume, 350 ml) and left atrial pressure was set at 1.5 mm Hg in all experiments (0 referenced at the hilus). The lungs were placed in a temperature-equilibrated chamber at 38° C, suspended freely from a force transducer for monitoring of organ weight. In parallel with the onset of artificial perfusion an air mixture of 80.5% N₂, 15% O₂ and 4.5% CO₂ were used for ventilation. A positive end-expiratory pressure of 1 cm H₂O was used throughout.

A/ Determination in Isolated Lungs by MIGET

The A/ distributions were determined by the multiple inert gas elimination technique, as described by Wagner and coworkers (20), which has been adapted to blood-free perfused rabbit lungs (21). An indication of acceptable quality of A/ distribution is a residual sum of squares (RSS) of 5.348 or less in half of the experimental runs (50th percentile) or 10.645 or smaller in 90% of the experimental runs (90th percentile) (22). In the present study 72.1% of residual sum of squares was less than 5.348 and 98.6% was less than 10.645.

Lung Lavage

For performing lung lavage, lungs were disconnected from the ventilator and warm isotonic saline (37.5 ° C) was instilled (total fluid volume, 13 ml/kg bw; instillation pressure, 20 cm H₂O). Fluid was then gently removed within 1 min, and ventilation was continued. During the lavage procedure, perfusion was not interrupted.

Aerosol Application

For aerosol delivery of saline or surfactant, the ultrasonic nebulizer was located between the ventilator and the isolated lung. The inspiratory tubing between ventilator and lung was heated to 40° C to prevent condensation. The aerosolized mass was calculated by weighing the nebulizer before and after aerosolization. Surfactant was suspended in isotonic saline at a concentration of 15 mg/ml, as pilot experiments indicated this concentration to be optimal for ultrasonic nebulizer efficacy. In the control with sham aerosolization, isotonic saline was nebulized. During aerosol delivery, undertaken over a 1-h period, the breathing pattern was altered to enhance deposition (frequency was increased from 11 ± 2 to 40 ± 2 breaths/min and tidal volume from 11 ml/kg to 14 ml/kg), both in the group with surfactant aerosolization and in the sham-aerosolized lungs. The ultrasonic nebulizer currently employed produced an aerosol with a mass median aerodynamic diameter (MMAD) of 4.5 µm with geometric standard deviation (GSD) of 2.3, as measured with a laser diffractometer (HELOS; Sympatec, Clausthal-Zellerfeld, Germany). Biophysical and biochemical properties of surfactant material aerosolized by the ultrasonic nebulizer remained unchanged as has been shown previously (13). For measurement of lung surfactant deposition, a laser photometer was located at the inlet of the tracheal tube, allowing breath-by-breath assessment of inhaled and exhaled aerosol mass (23).

Experimental Protocol

Studies were performed in six experimental groups (n = 6 lungs in each group). Control: no interventions were undertaken. After baseline A/ measurement, time was set at zero, and subsequent MIGET analysis was performed at 10, 40, 100, and 130 min. Control/ Inst.: after performing the time zero A/ measurement, 80 mg surfactant/kg bw (4 ml fluid/kg bw) were bolus-injected into the trachea, followed by a bolus injection of 9 ml air/kg bw. A/ measurements were performed as in the control lungs. Control/Aer.: after performing the MIGET analysis at times 0, 10, and 40 min, 36.4 mg surfactant/kg bw ( 2.5 ml fluid/kg bw) was nebulized by the ultrasonic device over a time period of 60 min. Further A/ measurements were performed at 100 and 130 min. Lung deposition of the nebulized material was 8.6 mg surfactant/kg bw in these experiments. Lav./Saline: after performing the time zero A/ measurement, lungs were lavaged as described. Between 40 and 100 min, sham nebulization of saline was performed (  2.6 ml fluid/kg bw). A/ measurements were performed as in the control lungs. Lav./Inst.: after performing the time zero A/ measurement, lungs were lavaged as described. At time 40 min, 80 mg surfactant/kg bw (4 ml fluid/kg bw) were bolus injected into the trachea, followed by a bolus injection of 9 ml air/kg bw. A/ measurements were performed as in the control lungs. Lav./Aer.: after performing the MIGET analysis at time zero, lungs were lavaged as described. Between 40 and 100 min, 35.2 mg surfactant/kg bw ( 2.4 ml fluid/kg bw) was nebulized by the ultrasonic device over a time period of 60 min. Further A/ measurements were performed at 100 and 130 min. Lung deposition of the nebulized material was 8.8 mg surfactant/kg bw in these experiments.

Statistics

Data are given as mean ± SE or as coefficient of variation (SD/mean, %). For analyzing statistical difference, Student's two-tailed t test for unpaired samples was performed. Differences were considered to be significant at a p value < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Control Lungs

After termination of the steady-state period, all lungs displayed mean pulmonary arterial pressures of 6 to 9 mm Hg (Table 1) and a ventilation pressure of 6 to 8 mm Hg. Baseline MIGET analysis revealed physiologic A/ distribution in all experiments. Unimodal narrow distribution of perfusion and ventilation to midrange A/ areas (0.1 < A/ < 10) was observed throughout, as indicated by a log standard deviation of perfusion (log SD ) of 0.35 to 0.50 and a log standard deviation of ventilation (log SD A) of 0.38 to 0.48. Shunt flow (A/ < 0.005) and perfusion to poorly ventilated regions (0.005 < A/ < 0.1) ranged below 2% and no perfusion of high A/ regions (10 < A/ < 100) was observed (data not shown in detail).

Instillation of Surfactant in Control Lungs (Control/Inst.)

As shown in Figure 1, instillation of 80 mg surfactant/kg bw resulted in a sharp rise of perfusion to low A/ areas, with a maximum of 12.8 ± 2.0% and a subsequent rapid decline, as well as a sustained increase in shunt flow (maximum, 6.4 ± 1.6%). This was accompanied by a marked increase in both log SD (1.4) and log SD A (1.41) (Table 1). Partial normalization of this A/ mismatch in the midrange A/ areas was observed over the subsequent 130-min observation period (example given in Figure 2). Surfactant instillation did not significantly change the pulmonary artery pressure (Table 1), whereas weight gain increased immediately after instillation to 10.2 ± 0.2 g and then remained largely constant. Ventilation pressure increased from 8.0 ± 0.5 to 12.5 ± 0.3 mm Hg in response to the instillation maneuver.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Intrapulmonary shunt flow and perfusion of low A/ areas in response to surfactant instillation in control lungs. At time zero, tracheal instillation of surfactant (80 mg/kg bw, corresponding to 4 ml fluid/kg bw) was undertaken. Mean ± SEM of six independent experiments are given; error bars are missing when falling into symbol. Low A/, percentage of perfusion of poorly ventilated areas (0.005 < A/ < 0.1); shunt, percentage of perfusion on nonventilated areas (A/ < 0.005).

View larger version (15K): [in this window] [in a new window]   Figure 2.   Ventilation-perfusion matching in response to surfactant instillation in a control rabbit lungs. The A/ distribution was assessed under baseline conditions ( top panel ), 10 min after instillation of 80 mg surfactant/kg bw (middle panel ) and at the end of experiment (bottom panel ).

Nebulization of Surfactant (Control/Aer.)

After nebulization of surfactant in control lungs (deposited amount, 8.6 mg/kg bw), A/ matching remained fully unchanged (log SD and log SD A given in Table 1). There was no increase in perfusion of nonventilated areas (shunt) or low ventilated areas (low A/). A total weight gain of ~ 7.6 g was noted, without any significant change in vascular perfusion pressure or ventilation pressure.

Lavage and Sham Nebulization of Saline (Lav./Saline)

As shown in Figure 3, the lavage maneuver provoked marked gas exchange abnormalities, characterized by a progressive increase in shunt flow to 21.9 ± 5.1% after 130 min. Concomitantly, the percentage of perfusion of areas with low A/ ratios was significantly increased (Figure 4), accompanied by a broadening of the perfusion distribution with an increase in log SD (p < 0.01) (Table 1). There was some minor increase in the dispersion of ventilation (log SD A). The lavage procedure provoked a very moderate rise in pulmonary artery pressure, from  8 to  9 mm Hg (NS) and a significant increase in ventilation pressure, from  7 to  14 mm Hg; p < 0.01). A total weight gain of 19.7 ± 2.1 g was monitored, which surpassed the fluid volume remaining within the lung because of the lavage procedure (~ 12 g).

View larger version (19K): [in this window] [in a new window]   Figure 3.   Intrapulmonary shunt flow in lavaged lungs subsequently undergoing surfactant instillation (Lav./Inst.), sham aerosolization (Lav./Saline), or nebulization of surfactant (Lav./Aer.). The instillation and nebulization maneuvers are indicated. Mean ± SEM of six independent experiments each are given; error bars are missing when falling into symbol. p < 0.001 as compared with the Lav./Saline group.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Perfusion of areas with low A/ ratios (0.005 < A/ < 0.1) in lavaged lungs subsequently undergoing surfactant instillation (Lav./Inst.), sham aerosolization (Lav./Saline) or nebulization of surfactant (Lav./Aer.). The instillation and nebulization maneuvers are indicated. Mean ± SEM of six independent experiments each are given; error bars are missing when falling into symbol. p < 0.05; p < 0.01 as compared with the Lav./Saline group; p < 0.01 as compared to the Lav./Aer. group.

Postlavage Surfactant Instillation (Lav./Inst.)

The "rescue" instillation of 80 mg surfactant/kg bw after performance of the lavage maneuver resulted in a significant decrease of shunt flow to a minimum of 3.5% (as compared with 22% in the absence of surfactant) (Figure 3), but there was a higher rate of perfusion in low A/ areas (maximum, 13.9% as compared with 5.3%) (Figure 4). An example of A/ distribution at presurfactant and postsurfactant instillation in lavaged lungs is given in Figure 5. The lavage-elicited shift to the left of perfusion (decrease in mean, not given in detail) and increase in log SD (Table 1) were not significantly improved by the subsequent surfactant instillation. This was also true for the dispersion of ventilation distribution (log SD A). The total weight gain in these lungs approximated 23 g, with no significant changes in pulmonary artery pressure. The rise in ventilation pressure was comparable to the lavaged lungs in the absence of surfactant instillation.

View larger version (13K): [in this window] [in a new window]   Figure 5.   Ventilation-perfusion matching in response to surfactant instillation post lavage (one example of the Lav./Inst. group). The A/ distribution was assessed under baseline conditions, 40 min after lavage (preinstillation), and at the end of experiment (postinstillation).

Postlavage Surfactant Nebulization (Lav./Aer.)

Postlavage nebulization of 35.2 mg surfactant/kg bw, resulting in a lung deposition of 8.8 mg/kg bw, caused a significant reduction of shunt flow (maximum, 7.4%) as compared with lavaged lungs undergoing sham aerosolization (Figure 3). In addition, perfusion of low A/ areas was markedly decreased (maximum  2%) (Figure 4), and a more narrow distribution of perfusion was noted, indicated by a decrease in log SD in response to surfactant nebulization (p < 0.05) (Table 1). An example of A/ distribution at presurfactant and postsurfactant nebulization in lavaged lungs is given in Figure 6. Pulmonary artery pressure remained unchanged, and 130 min after lavage a weight gain of 17.1 ± 2.0 g was measured. The ventilation pressure was not significantly different from sham-nebulized lungs.

View larger version (15K): [in this window] [in a new window]   Figure 6.   Ventilation-perfusion matching in response to surfactant nebulization post lavage (one example of the Lav./Aer. group). The A/ distribution was assessed under baseline conditions, 40 min after lavage (prenebulization), and at the end of experiment (postnebulization).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study compared the effect of an ultrasonically manufactured surfactant aerosol, deposited at low dosage in the lung tissue, with that of a tracheal-instilled "conventional" surfactant dose in an acute lung injury model provoked by lung lavage. This procedure disrupted the physiological narrow unimodal distribution of ventilation and perfusion, with the appearance of shunt flow, perfusion of low A/ areas, and severe A/ mismatch. Rescue application of surfactant via tracheal injection dramatically reduced the shunt flow, but at the same time increased the perfusion of low A/ areas. Correspondingly, the appearance of low A/ areas and A/ mismatch were also provoked by tracheal surfactant instillation in control lungs. In contrast, ultrasonic nebulization of surfactant did not interfer with physiological A/ matching in control lungs, and when the surfactant aerosol was administered in lavaged lung, shunt flow, perfusion of low A/ areas, and A/ matching all improved. Low doses of ultrasonically delivered surfactant may thus even be superior to "conventional" doses of tracheal-instilled surfactant in improving gas exchange in a homogeneous model of acute lung injury.

In ARDS, several surfactant inhibitory mechanisms may be encountered (1). These include decomposition of lipophilic and hydrophilic compounds, accelerated conversion of large to small aggregates, and inhibition of surfactant function by plasma protein leakage. We focused our interest on the gas exchange abnormalities in a highly reproducible model of lung injury in which surfactant abnormalities represent the predominant underlying event. Therefore, we intended to be independent of extrapulmonary mechanisms, which might additionally effect surfactant function in a less controlled fashion. Rather than designing an as much realistic model of ARDS as possible, we thus chose isolated perfused lungs to compare the efficacy of surfactant instillation with that of an ultrasonic nebulization technique. Saline lavage techniques of the entire lung are commonly used as ARDS models (24), characterized by atelectasis, intraalveolar protein leak and severe hypoxemia (25, 27, 28). The present assessment of A/ distribution in this model emphasizes serious abnormalities, with a predominance of shunt-flow, appearance of low A/ areas, and profound ventilation-perfusion mismatch in the midrange A/ regions, which are reminiscent of the gas exchange disturbances reported for patients suffering from ARDS (29, 30). Excellent reproducibility of these abnormalities is indicated by the relatively low SEM bars of all variables assessed. Hemodynamics were largely unchanged, but the loss of compliance demanded increased peak inflation pressures because of the mode of volume-controlled ventilation.

On the basis of previous experimental studies as well as theoretical considerations (8), both favoring tracheal bolus application of surfactant as compared with tracheal infusion of the liquid material for more homogeneous distribution within the lung, a rapid bolus injection technique was employed for tracheal surfactant administration in the current study. When administering a "conventional" surfactant dose of 80 mg/kg bw in control lungs, in an adequate total volume of 4 ml/kg bw, marked gas exchange abnormalities were created. A sharp rise in perfusion to low A/ areas, accompanied by some minor shunt flow, and the appearance of major ventilation-perfusion mismatch in the midrange A/ regions (enhanced log SD and log SD A) are all most readily explained by ventilation inhomogeneities because of airway occlusion by the viscous liquid material. This view is also supported by the immediate increase in peak inflation pressure. Secondary changes in perfusion distribution, in response to the altered pattern of gas flow distribution may further add to the A/ profile provoked by surfactant instillation in the normal lungs.

When instilled in lungs that have undergone the lavage procedure, the dose of 80 mg Alveofact/kg bw caused a rapid and highly significant decline in shunt flow, as anticipated from preceding experimental studies with tracheal surfactant instillation in models of acute lung injury (31, 32). However, perfusion of low A/ areas increased rather than decreased, and ventilation-perfusion mismatch in the midrange A/ regions was not improved. These events are all reminiscent of the surfactant bolus-induced alterations in the control lungs. Superposition of beneficial effects caused by opening of atelectatic regions and disadvantageous effects caused by airway occlusion phenomena with impact on gas flow distribution are thus suggested by the MIGET data. Similar results, i.e., decrease in shunt flow but increase in the perfusion of low A/ areas, were previously obtained when assessing gas exchange at presurfactant and postsurfactant administration in patients with ARDS (4).

The surfactant aerosol application in the current study was performed by means of ultrasonic nebulization, with a device producing particles with a mass median aerodynamic diameter (MMAD) of 4.5 µm and a geometric standard deviation of 2.5. As previously shown, biophysical and biochemical properties of ultrasonically nebulized surfactant are unchanged (13). The aerosolization technique resulted in lung surfactant deposition of nearly 9 mg/kg bw within 1 h, thus significantly surpassing the range of 4 to 5 mg surfactant/kg bw deposited within 3 h by use of jet nebulization (14). Interestingly, no difference in lung surfactant deposition was noted between control and diseased lungs, whereas in a preceding study in preterm lambs a markedly reduced deposition of aerosolized surfactant material in low-compliance as compared with high-compliance lungs was observed (18). In the control lung group undergoing ultrasonic nebulization for 1 h, no disadvantageous effects of this route of transtracheal surfactant administration were noted, but physiological ventilation-perfusion matching was fully maintained. In lungs injured by preceding lavage, a marked reduction of shunt flow, not significantly different from that in response to tracheal bolus application of a nearly tenfold higher dose of surfactant, was again achieved. In contrast to the bolus application, however, perfusion of low A/ areas was not increased but significantly declined upon surfactant nebulization. The same was true for ventilation-perfusion matching in the midrange A/ areas: a more narrow distribution of perfusion and ventilation was achieved, as shown in Figure 6 and indicated by the lower numbers of log SD and log SD A. Transtracheal delivery of ultrasonically nebulized surfactant thus evidently caused recruitment of atelectatic regions and thereby reduction of shunt flow, without any evidence of bronchial obstruction and related provocation of ventilation inhomogeneities. The fact that the surfactant quantities required for significant reduction of shunt flow ranged approximately one order of magnitude below the quantities "conventionally" (also in this study) applied by tracheal surfactant instillation is well in line with the preceding studies of Lewis and coworkers (14), employing jet aerosolization in comparison with surfactant bolus injection in different models of acute lung injury.

What are the reasons for the obviously higher efficacy of aerosolized surfactant as compared with tracheal-instilled material in improving gas exchange, along with preservation of physiological ventilation-perfusion matching in healthy lungs? Theoretically, liquid surfactant material deposited in the large airways should rapidly spread to the alveolar surface, representing the predominant surface area, because of excellent adsorption characteristics and lateral spreading facilities of natural surfactant material known from a large body of in vitro studies (9, 10). The MIGET measurements in the healthy lungs undergoing surfactant administration do, however, question this concept, but rather favor the view that large quantities of liquid surfactant material are retained in the airways for at least the current observation period of 2.5 h, thus limiting the supply of the alveolar surface and provoking serious ventilation inhomogeneities. Direct peripheral deposition of aerosolized surfactant may well overcome these shortcomings, and the A/ data in both the healthy and the diseased lungs in the current study clearly support this view.

Our investigation is limited in that a model of homogeneous lung injury was employed, which may be more advantageous for surfactant aerosol approaches than lungs with nonuniform patterns of lung lesions. Moreover, as being transported via gas flow, nebulized surfactant must be expected to be predominantly deposited in well-ventilated alveolar spaces, and not in atelectatic regions representing the main target areas ("aerosolization goes to parts of the lungs where it is least needed"). However, at least in the present model, surfactant deposited in the lung periphery by the nebulization technique did evidently spread to the atelectatic areas, as true shunt flow measured by MIGET technique was highly significantly reduced. Further developments might combine aerosolization techniques with short-term mechanical lung recruitment ("open up") maneuvers (increase of inspiratory and expiratory pressures), to profit from better gas flow access to the diseased regions for surfactant delivery. Such a concept, supported by recent studies in the combination of volume recruitment and surfactant instillation, will then demand nebulization techniques transporting as much surfactant material per time as possible. The present finding that ultrasonic aerosolization is well suited for efficient surfactant delivery to the lung periphery and exerting advantageous effects on ventilation-perfusion matching may lend credit to this approach.